After noticing that locations polygons are reduced in their edges number to reduce load on iNaturalist servers, resulting in observations that are considered to be in the wrong location boundary (eg. country) while it’s waypoint is accurate, I was thinking what could make querying more efficient and allow for full edges polygons.
But then I’ve realized (with the help of a friend) that computing the relation of an observation waypoint to places boundaries (polygons) happens only once at observation creation (or update), so it doesn’t need to be recomputed in every page load, what also makes this not so heavy in terms of server load.
Can you please elaborate on this topic?
Why not using standard openstreetmap polygons (which are very accurate and updated by the community of openstreetmap) instead of using inaccurate polygons for places boundaries which make observations relate to wrong boundaries, and also cost more curator work to fix and update locations from time to time?
Thank you. I understand that this is not a technical problem (which can be solved), but that there is no best source of truth (haven’t looked at this myself).. I also recall that on some countries there is an ongoing process of making boundaries more accurate using GPS. Can iNaturalist be automatically updated with this process results?
i would think that updating iNaturalist’s standard places is not a small task. if you search the forum for World Tour, you’ll see how they sort of went about it last time. that was almost 7 years ago. i’d be surprised if they considered tackling the standard places again before 10 years had passed.
if they shift to a boundary data set that is updated frequently, it’s possible that they might be able to update boundries more frequently after that, but i’m not sure how automated that sort of process can ever be, and that initial transition will likely take significant human effort.
I understand that the world tour deals with observations more than boundaries. If you can find the discussion of the process of updating boundaries that was done 7 years ago that would be great. The borders of Israel haven’t changed for many years though, if there’s a way to update Israel’s boundaries that would be great. I also have level 2 boundaries for Israel, I will speak with astra_the_dragon on that.
I understand that updating standard places is not a small task, it’s not just replacing json files because places are splited and added or also united or removed.
Maybe updating boundaries should not be automatic, but it seems that there are many discussions in the Forum and flags in iNaturalist regarding boundaries (not just in level 1), so I think that denying the request of fixing boundaries can’t be a policy either. If that’s not an infrastructure technical issue I can continue the discussion in the flag I’ve opened on Israel, and I would be glad for your word there either.
Anyway, if accurate polygons are available, I would suggest using them to fix what can be fixed and open a discussion when a conflict between two sources of truth arises..
if you’re limiting the scope to this, it might be easier to convince iNat staff to do this, especially if you get support from the iNaturalist Israel node administrators and/or significant scientific organizations in the region. (adding new boundaries should be easier to do than modifying existing boundaries.)
but i have no idea how iNat staff set their work priorities.
astra_the_dragon did bring up the issue of level 2 boundaries back in the day, and if that never went anywhere in 6 years, i’m not sure if the roadblock would have been related to lack of a good source, lack of interest / prioritization, or maybe even venturing too closely to hot-button politics.